Traditionally, linear phenotypic selection analyses and reciprocal transplants have been used to examine natural selection and local adaptation, respectively. These techniques, while easily comparable among studies, can potentially miss the selection and evolution experienced by these populations. In this study, we experimentally introduced a population of Brassica rapa from Southern California to New York in order to examine selection, and the potential for rapid evolution, of phenotypic traits directly following introduction to a novel environment. Along with traditional phenotypic selection analyses, we utilized a variety of models that examine a different aspects of fitness functions, including asymmetry and thresholds, in order to get a better understanding of the selection pressures experienced by the introduced population during the first three years following introduction. We also examined selection on different fitness components, as well as comprehensive estimations of fitness, in order to examine selection on different stages of the life cycle. In order to examine evolution since introduction, we utilized the resurrection approach of comparing ancestors and descendants in a common environment. We examined evolution of a variety of different morphological, phenological, and physiological adult traits, fitness proxies, and seed traits to examine evolution in multiple stages of the life cycle.
Results/Conclusions
Classic phenotypic selection analyses demonstrate selection for larger size and earlier flowering time that varied in strength, but not direction, during the first three years following introduction. Alternatives estimations of selection, including exponential decay and piecewise linear models, suggest that linear differentials and gradients underestimate the strength of selection on many, but not all, phenotypic traits. Comprehensive estimations of fitness increase the fit of the model to the data, which suggests that they provide a better estimation of total fitness. However, selection on different fitness components mimics selection on seed pod count for virtually all traits. The results of the resurrection experiments demonstrate evolution of a variety of adult traits, including the evolution of smaller overall size, smaller leaves, earlier flowering time, and shorter duration of flowering. However, there was no observed evolution in any seed traits that were examined, including emergence time in the spring or adaptive dormancy in the fall. This suggests that adult traits may to rapidly evolve more easily, or may be less plastic, than seed traits in response to novel conditions. Overall, these results demonstrate the abundance of information that can be collected when going beyond traditional estimations of selection and evolution.